The Cognitive Psychology of EHR/EMR Usability and Workflow

Short Link: http://j.mp/7fbVl1

Which targets are easier to hit quickly, accurately, and repeatedly? Small checkboxes or large buttons?

If this was an interactive Flash demo, I’d give you some instructions about what to click in what order and then present back to you some speed and error statistics. Since it is not a Flash demo–yet–at the end of this post, I provide a link to a more sophisticated online demo of Fitts’s Law that does exactly that.

For one year as a graduate student in Industrial Engineering, I worked as a programmer at the Coordinated Science Laboratory at the University of Illinois in Champaign-Urbana on an aviation human factors project studying pilot error. (Workflow related, perhaps worth a post. The other year I worked on a computer simulation of patient workflow through the student health center. Also workflow related and perhaps worth a post.) Professors Bill Rouse and Christopher Wickens (Aviation Research Lab) and their graduate students met regularly to present and discuss their research. There is where I first encountered two psychological laws that are highly relevant to high-productivity heads-down data input and order entry, which means they are highly relevant to pediatric EMR workflow systems.

Fitts’s Law: “The time required to rapidly move to a target area is a function of the distance to and the size of the target.” (Wiki article on Fitts’s Law. Paul Fitts was an Air Force psychologist.)

Hick’s Law: “The more choices you have to choose from, the longer it takes for you to make a decision.” (There’s a Wiki article, but this one on a Taekwondo site is a lot more fun. At one time or another who hasn’t wanted to “strike or break with a fist” some badly designed piece of software?)

In human factors speak, an EHR button or checkbox or tab is a “target.” Why do toys and senior-friendly products have just a few large buttons? Because these large targets are easier to “acquire” (press, click, touch, etc.) quickly and accurately than smaller ones.

Applied to EHR/EMR user interfaces, Fitts’s and Hick’s laws translate into “lots of small buttons all over the place are incredibly time consuming, frustrating, and error prone.” This is why the EMR workflow systems have just a few large consistently placed buttons on each screen.

How, might you ask, can EMR workflow system present *enough* buttons to a physician so that they can enter all the data and orders that they need? Funny you should ask…instead of just a few big screens containing many small buttons and checkboxes and so on, we spread larger buttons (and no checkboxes, not a one) across many screens. How, might you ask, are you expected to navigate to the right screen at the right time in order to click on the right button? Again, funny you should ask…for each specific situation (well child visit, sick child visit, vaccination, etc.) we present the right screens in the right sequence back to you in a way that mirrors the natural order of the tasks you need to accomplish. That is what a workflow management system does (among many other valuable services). It is intrinsic to its nature. It does the scut work of navigating to the next screen for you.

One physician who uses the an EMR workflow system dismisses non-workflow automated pediatric EMRs as “hunt-and-peck” EMRs, because users have to hunt and peck (click) in order to figuratively drag the EMR through a patient encounter. He calls his EMR workflow system an “anticipatory EMR” because, like a good OR nurse,  it anticipates what you want to do next and hands you the right data input or order entry tool to accomplish it quickly and without error.

Another pediatrician who uses an EMR Workflow System was a pre-med music major. It is rather uncanny to watch him “play” his EMR. Picklists are large wide buttons reminiscent of piano keys (rotated 90 degrees). He sort of turns his hand so that one finger strikes one key on one screen while another finger strikes another key on the next screen and so forth in quick succession. Data and order entry in a pediatric EMR workflow system relies on cognitive and motor skills more similar to that of a musician than a knowledge worker.

I’ve been at this for a long time. In 1991 I wrote the following comment about EMR usability and workflow (originally published in the Journal of Medical Practice Management, but reprinted here, page 11). In light of slow EMR adoption rates that are due in large part to slow EMR data input and order entry, the comment is even more relevant today:

The choice of touch screen technology and large icons deserves some comment. A major motivation for using a structured data entry approach is not just to obtain structured data but also to increase the speed of data entry. Fitts’s Law [5] is a mathematical model of time to hit a target. It basically says that larger targets are easier, and faster, to hit. Fitts’s Law seems obvious, but it is often ignored when designing electronic patient record screens because the larger the average icon, button, scrollbar, etc., the fewer such objects can be placed on a single screen. A natural inclination is to display as much information as possible; EPR screens are thus often crowded with hard-to-hit targets, slowing the user rates of data entry and increasing associated error rates.

Fitts’s Law, in conjunction with constrained screen “real estate,” suggests use of a few, large user-selectable targets. Displaying fewer rather than many selectable items tends to increase the number of navigational steps, unless some approach is used–such as a workflow system–that automatically and intelligently presents only the right structured data entry screens.

In our opinion, the combination of structured data entry, workflow automation, and screens designed for touch screen interaction optimally reduces inherent tradeoffs between information utility and system usability on one hand, and speed and accuracy of data entry on the other. Successful application of touch screen technology requires that only a few, but necessary, selectable items be presented to the user in each screen. Moreover, workflow, by reducing cognitive work of navigating a complex system, makes such structured data entry more usable.

I couldn’t have put it better myself. (I do know that quoting oneself is pretty darn self indulgent, but, hey, this is a blog.)

By the way, here is that really cool interactive flash demo of Fitts’s Law that I referred to earlier (and a screen shot from that demo).

I’m thinking about doing something similar in Flash with respect to simulated data input and order entry using a simplified EMR user interface. There would be two versions: (1) A single screen with a hundred small buttons in a ten-by-ten grid, and (2) a succession of ten screens with ten large buttons apiece.  Similar to the interactive demo of Fitts’s Law, buttons would flash and the user would try to click the indicated button as quickly as possible (“cued target acquisition” in human factors parlance). After clicking ten buttons the total time and error rates would be displayed. (Dear medical informatics student, this sounds like a fun project, send me a link when you are done and I’ll post it.)

With apologies to the human factors community:

There once was a pilot named Fitts:
“Large targets are more easily hit.”
    His words became law:
    “Small is a flaw!”
And target for my doggerel wit.

(Doesn’t sound too bad if you slightly deemphasize the “s” in “Fitts”.)

P.S. There is more about EMR usability and workflow in a previous post: “Pediatric EMR Usability: Natural, Consistent, Relevant, Supportive, Flexible Workflow.”